allure-agent-mode-setup
Pass
Audited by Gen Agent Trust Hub on Apr 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill directs the agent to ingest and reason about data from external files such as 'index.md' and JSONL manifests, which represents an indirect prompt injection attack surface.
- Ingestion points: manifest/run.json, manifest/tests.jsonl, and manifest/findings.jsonl.
- Boundary markers: No explicit delimiters or instructions to ignore embedded instructions are present in the guide templates.
- Capability inventory: The agent uses the content of these files to determine root cause, evaluate regression safety, and make final quality statements.
- Sanitization: No explicit sanitization or schema validation of external manifest content is described.
- [COMMAND_EXECUTION]: The skill contains instructions for the agent to execute shell commands including 'allure agent' subcommands for state management and reruns.
- [EXTERNAL_DOWNLOADS]: The skill suggests running 'npx allure agent', which involves fetching and executing the 'allure' package from the public npm registry.
Audit Metadata